Evidence at 2024 World Conference on Lung Cancer cites chest X-ray AI’s potential to detect lung nodules early

This study aimed to investigate the benefit of artificial intelligence (AI) chest X-rays in detecting missed lung cancer diagnoses in community-based cancer centres where most of the chest X-ray interpretation is done by general practitioners.

Methods: We retrospectively reviewed the chest X-ray image database of newly diagnosed lung cancer patients from 1 January to 31 December 2022 at Phrapokklao Cancer Center of Excellence. A total of 197 patients were screened, the eligible 77 patients who obtained more than six months before lung cancer (LC) diagnosis. These images were evaluated by CXR Qure.AI software to identify abnormal nodules or perihilar thickening for more than six months before a definite diagnosis of lung cancer, defined as missed lung cancer (MLC) diagnoses. We assessed the median time of MLC diagnosis.

Results: Among the 77 patients analysed, fourteen patients (18.18%) were found to have MLC diagnoses. The mean duration of MLC diagnosis was 32.3 months with a 95% CI from 20.6 to 44.08 months, with a maximum of 101 months and a minimum of 8.2 months. Of 14 MLC patients, six patients (42.9%) were diagnosed with stage IV lung cancer, while three patients (21.1%) were diagnosed with stage III lung cancer.Conclusions: This study's findings show the potential of AI technology to facilitate the detection of lung cancer from chest X-ray images.

The poster presentation of this took place at International Association for the Study of Lung Cancer (IASLC) 2024 / World Conference on Lung Cancer in San Diego 2024.